Hi Flink devs, What do you think about releasing Flink 1.8.1 soon?
We already have some critical fixes in the release-1.8 branch: - FLINK-12301: Fix ScalaCaseClassSerializer to support value types - FLINK-12342: Add fast-heartbeat-delay yarn config for jobs with a large number of containers - FLINK-12260: Slot allocation failure by task manager registration timeout and race - FLINK-10941: Keep slots which contain unconsumed result partitions - FLINK-12042: Fix RocksDBStateBackend's mistaken usage of default filesystem Furthermore, I think the following 3 Blocker issues and 2 Critical issues should be merged before 1.8.1 release. [Blocker] - FLINK-11987 Kafka producer occasionally throws NullpointerException - FLINK-12296 Data loss silently in RocksDBStateBackend when more than one operator(has states) chained in a single task - FLINK-12297 Clean the closure for OutputTags in PatternStream [Critical] - FLINK-10455 Potential Kafka producer leak in case of failures - FLINK-11059 JobMaster may continue using an invalid slot if releasing idle slot meet a timeout I know that the reason for the proposal release 1.8.1 is there is already a very important bug that is fixed in 1.8-SNAPSHOT. And for the reason of the proposal that merges the above 3 Blocker issues and 2 Critical issues, that because those issues not only for the 1.8.1 but also very important for the 1.9.0 release which the deadline already mentioned in mailing thread by Gordon. So I think it's better to deserve to do this effort in the 1.8.1 release process to get as much time and quality assurance as possible for the 1.9.0 release. Please let me know what you think? Cheers, Jincheng
